COOM: A Game Benchmark for Continual Reinforcement Learning.
Tristan TomilinMeng FangYudi ZhangMykola PechenizkiyPublished in: NeurIPS (2023)
Keyphrases
- reinforcement learning
- game theory
- function approximation
- video games
- game playing
- game theoretic
- educational games
- state space
- optimal policy
- comparative analysis
- learning process
- multi agent reinforcement learning
- learning capabilities
- game design
- game based learning
- reinforcement learning algorithms
- robotic control
- nash equilibria
- two player games
- benchmark suite
- real world
- multi agent
- computer games
- supervised learning
- machine learning
- reinforcement learning agents
- temporal difference
- markov decision processes
- monte carlo tree search
- temporal difference learning
- virtual world
- dynamic programming
- policy search
- game players
- imperfect information
- game play
- online game
- monte carlo
- transfer learning
- action selection
- model free
- nash equilibrium
- serious games